Skip to main content

Verify Cluster Connection

Overview​

After installing the OpsWorker Kubernetes Agent, confirm that your cluster is connected and data collection is working. Verification is automatic: the portal polls the agent status and updates the cluster state on its own. There is no manual test button.

Validation Steps​

1. Check Cluster Status in Portal​

Navigate to your cluster in the OpsWorker portal. The status reflects the agent's onboarding state:

StatusMeaningAction
PENDINGAgent not yet installed or not yet connectedInstall the agent or wait a minute
SUCCESSAgent connected (shown as connected/online in the UI)Ready to go
FAILUREOnboarding failedCheck the agent pods and logs below
TIMEOUTAgent did not report in timeCheck connectivity and reinstall if needed

2. Check Agent Pods​

Verify the agent pods are running in your cluster:

kubectl get pods -n opsworker-agent

A default install creates three pods (the agent, kubernetes-mcp-server, and grafana-mcp):

NAME                                          READY   STATUS    RESTARTS   AGE
opsworker-agent-xxxxxxxxxx-xxxxx              1/1     Running   0          2m
opsworker-agent-kubernetes-mcp-server-...    1/1     Running   0          2m
opsworker-agent-grafana-mcp-...              1/1     Running   0          2m

3. Check Agent Logs​

Review the agent logs for a successful connection:

kubectl logs -n opsworker-agent -l app.kubernetes.io/name=opsworker-agent

Look for messages indicating successful SQS connection and polling.

4. Confirm End to End​

To see a full investigation run, either:

  • Wait for a matching alert from your connected alerting platform. Alerts that match an alert rule trigger an investigation automatically.
  • Use AI Chat in the portal to ask a question about your cluster (e.g., "Are there any pods in CrashLoopBackOff?"). The agent answers in real time, which confirms the path from portal to cluster and back.

Troubleshooting​

If your cluster does not reach SUCCESS:

  • Pods not running: check kubectl describe pod -n opsworker-agent -l app.kubernetes.io/name=opsworker-agent for scheduling or image pull issues
  • Outbound connectivity: verify the agent can reach AWS SQS, AWS STS, api.opsworker.ai, and onboarding.ow-saas-prod-ext.opsworker.ai on port 443
  • Network policies: check whether any NetworkPolicy resources block outbound traffic from the opsworker-agent namespace
  • Credentials: confirm the install used the exact command from the portal (both clusterToken and apiKey are required and auto-filled)
  • Proxy: if behind a corporate proxy, set HTTPS_PROXY/NO_PROXY via the chart's env list

See Agent Connectivity Troubleshooting for detailed guidance.

Next Steps​